#asahi-dev · 2026-01-11 23:35
svenSPTM was only rolled out for M3 and M4, wasn't it?#asahi-dev · 2026-01-11 23:20
sventhe ipsws contain the adt templates though and you can extract them, they should have stuff like MMIO addresses#asahi-dev · 2026-01-11 23:20
svenfyi: be careful when posting adt files dumped with m1n1, they contain stuff like bluetooth and wifi keys in not-so-obvious places#asahi-dev · 2026-01-11 23:16
svenoh no, absolutely not#asahi-dev · 2026-01-11 23:14
svendon't think anyone's going to complain if you make the code more accessible :-)#asahi-dev · 2026-01-11 23:13
svensure!#asahi-dev · 2026-01-11 23:12
svennow if that works i can only think of hardware differences and then we're getting into really obscure stuff. like maybe m3 speculates harder and we're missing barriers somewhere? i dunno#asahi-dev · 2026-01-11 23:08
svenif it was some new feature xnu exercises we'd also see all that crap with macos 26+ on m1/2. has anyone ever tried to run that in the hv?#asahi-dev · 2026-01-11 23:07
svenso i'm still wondering what would cause all that brokeness on m3 but not on m1/2#asahi-dev · 2026-01-11 22:58
svenwe used to boot linux and even macos in the hv without mcc.c iirc, it just makes things faster#asahi-dev · 2026-01-11 21:40
sventhe linux driver definitely doesn't implement any of that though#asahi-dev · 2026-01-11 21:39
svenso maybe we know how it works#asahi-dev · 2026-01-11 21:39
svenhuh.. actually, looks like it's implemented for the secondary cores at least#asahi-dev · 2026-01-11 21:38
sventhat maybe enters the deep sleep state, but i haven't seen it exit from that one and show up at RVBAR again#asahi-dev · 2026-01-11 14:53
svenworks here as well#asahi-dev · 2026-01-11 14:32
svenyay! also el1_call fix PR would be nice ;)#asahi-dev · 2026-01-11 14:19
svenhttps://f.svpe.de/a330911e3c402e201618adc0b080ed4a44b0b13e981198cf4d63f6726443e3cc_0001-power-hibernate-Disable-hibernation-on-Apple-Silicon.patch only compile tested#asahi-dev · 2026-01-11 14:15
svenof_machine_is_compatible is even better#asahi-dev · 2026-01-11 14:14
svenof_machine_compatible_match against apple,arm-platform should work i guess#asahi-dev · 2026-01-11 14:11
svenat least downstream we could just add a hack to hibernation_available that makes it always return false on apple silicon#asahi-dev · 2026-01-11 14:09
sventhat would at least prevent all devices from being put to sleep but i'm not sure if that would stop the entry->exit loop#asahi-dev · 2026-01-11 13:53
svenpower management was set to hybrid sleep which is what triggered hibernation i think#asahi-dev · 2026-01-11 13:52
svenhttps://f.svpe.de/eaaca1067e33b62e3e069ab146cb2a2a7d66b4b0d5158a05762863ff0f0e9c29_asahi-diagnose-20260111-093152.txt#asahi-dev · 2026-01-11 13:50
svenand then there's a bug in some driver that only rarely happens and gets hit when we try to constantly hibernate#asahi-dev · 2026-01-11 13:49
svenmaybe userland doesn't notice about that failure and then immediately tries again?#asahi-dev · 2026-01-11 13:47
svenand then immediately tries again#asahi-dev · 2026-01-11 13:47
svenand then it brings everything back up#asahi-dev · 2026-01-11 13:47
svenJan 11 01:34:49 sven-macbookair kernel: Error taking CPU7 down: -95 <-- i think that's what stops it in the end#asahi-dev · 2026-01-11 13:47
svenit's been trying to hibernate for ~2 hours last night before it finally died somewhere i think#asahi-dev · 2026-01-11 09:22
svenyes, that's obvious. what i want to figure out is how the kernel determines if hibernation is supported without a boot arg#asahi-dev · 2026-01-11 09:02
svenbut the issue still happens even with writing to it only?#asahi-dev · 2026-01-11 09:01
svenhuh#asahi-dev · 2026-01-11 08:54
svendoes it just not come up?#asahi-dev · 2026-01-11 08:54
svendo you know why nvme doesn't work on m3?#asahi-dev · 2026-01-11 08:50
svenok, will try to grab one later#asahi-dev · 2026-01-11 08:50
svenoh right, do you still need that nvme trace?#asahi-dev · 2026-01-11 08:49
sveni just wonder what makes the system think it can do that#asahi-dev · 2026-01-11 08:49
svenyeah, not sure either. but dmesg looks like it tried to hibernate#asahi-dev · 2026-01-11 08:49
svenwhich we just don't support on these machines#asahi-dev · 2026-01-11 08:48
sventhat sounds like my issue, i suspect that happens when it tries to enter hibernate for some reason#asahi-dev · 2026-01-11 08:47
svenyes, it's definitely unrelated#asahi-dev · 2026-01-11 08:45
svendo i need to mention again that that branch is meant for developers who can debug issues?#asahi-dev · 2026-01-11 08:45
sven*sigh*#asahi-dev · 2026-01-11 08:37
svenwe shouldn't ever try to enter that tbh#asahi-dev · 2026-01-11 08:35
svenso 100%#asahi-dev · 2026-01-11 08:35
svenit's connected to power#asahi-dev · 2026-01-11 08:33
svennot sure why it even tries to hibernate since we don't really support that#asahi-dev · 2026-01-11 08:32
svenlast dmesg entry, has anyone seen that before?#asahi-dev · 2026-01-11 08:32
svenJan 11 02:28:24 sven-macbookair kernel: PM: hibernation: hibernation entry